Weekly news and teardowns on trending startup companies
The most relevant Twitter threads of the week
Product of the week brought to you by ProductHunt
Detailed stats on the companies that raised venture capital
This week, you couldn’t scroll through social media without seeing your friends sharing what they’d look like with different eye colors, or as astronauts or aliens. Some images were uncanny, concerningly precise, and others seemed more a joke than anything else. As these images littered our social media feeds, many wanted to know where they came from. The answer, from the outside, is innocent: it's an app, one of many, but deep down, it's much more frightening.
These remarkable images come from a computer, yet simultaneously come from the entire internet all at once. They're the result of artificial intelligence, something we've created and don't yet fully understand. The problem is that it has proliferated past our control and brought forth a barrage of ethical and legal questions hidden behind entertaining images.
One company that’s primarily responsible for the latest surge in AI imagery is Lensa.ai. Its use grew exponentially in late 2022, but the product isn't new. It's the brainchild of Prism Labs, which created it in 2018. However, it's only recently that the app has gained impressive traction. After all, the app had 4 million downloads in the first five days of December. So, what's the reason for this colossal surge? It's all due to avatars. No, not the movie, but Lensa's new Magic Avatars lineup of AI-enhanced selfies takes the user to fantastic lands and exploits different artistic styles.
At first, most users were excited. All it took was uploading eight to ten selfies into a massive database that would process them and transform them into avatars in no time. The end result was uncanny. There's no denying it. It was the first time we could actually see ourselves as practically anything we could imagine, but then many saw past the hype. Questions surged, and answers have yet to arrive.
As users saw themselves in the pictures, many were baffled at the extreme perfection of the end result. Anything from wrinkles to scars disappeared. Also, features such as chin and eye shape, nose size, and even the smile changed. Instead came into view perfectly smooth skin, which, by the way, also changed in hue and color, consistently lightening the tone.
AI creates uncanny selfies but takes liberties with them, shaping them more to oversexualized standards. So, it seems that we've also taught AI racism and sexism. The concerns don’t end there. What if this AI does the same with children? There's no limiting factor to age; unfortunately, it didn't take very long for the world to see that it did. It's frightening, but it's where we are headed.
Lensa is only one of many AI programs solely designed for generating content. In recent times, technology has advanced so much that we can create images, write text and even compose music with a single prompt. From that initial idea, we let AI take over and it begins millions of iterations, using previous content to generate a new flow of information. It sounds excellent, but it's at this point that AI has failed, and there's no way to fix it because it's in its essence.
AI needs previously existing content to create its own, and all these platforms share one thing in common. They can scour the web and copy millions (if not more) of images, songs, and texts. From there, it's just a matter of iterating until they come up with a final product that most of us love - except for those who created the base content in the first place.
On top of that, AI is so efficient that you, as an artist, will have no idea when it's using your art as a basis for its magical selfies. It's only when someone shares an image that bears your unique artistic style that you’ll know. It's then that many artists have asked if there's something they can do.
AI-generated content presents one significant dilemma concerning copyrighted material. The platforms are making millions thanks to subscription services or one-time payments. Yet, the artists that helped inspire the styles get nothing. The reason is that there's nothing to stop them. Now, I don't say that all AI platforms are taking advantage of this, but some might. Right now, most countries lack regulations around AI-generated content. So, it's almost impossible to pin ownership and blame anyone who uses it. So, here's where I go back to the essential process by which AI works: the initial human prompt.
Ever since we discovered art, our creativity and drive have defined it as art. However, it's the human authorship that makes it unique. Yes, even before this AI dilemma, there have been considerable discussions surrounding what's copying and what isn't. Art will always convey this debate, but AI-generated art revolutionizes and makes it more complex because it removes the human side of the equation.
Before writing this, I thought practically anyone could create a prompt and tell the AI platforms what to do. So, that end product would belong to the person who prompted the platform, but I'm wrong. At least, that's what many Governments believe. AI-generated artwork is not copyrightable in the US because it lacks human authorship. That image belongs to no one and everyone at the same time. Though these images can carry your style as an artist, there's currently nothing you can do.
Then, there's another question roaming in my mind. What if AI creates an image that looks exactly like you without your permission?
These platforms are scanning millions of images that exist on the web, and it's not just art; they need human references as well. Plus, with many people uploading their selfies to these databases, an AI can likely generate a random image that looks like you, which someone can use for commercial purposes because nobody can stop them. There's no regulation when it comes to AI-generated art. Soon, you might find yourself on a billboard ad, or a pop-up on the web, and there's nothing you can do about it.
I won't deny it: those selfies are tempting, and I want to use the platform, but at the end of it all, Lensa is just a fad, so it's unfair to chastize only this one. It's not the only app doing this. The thousands of platforms out there already will evolve to unfathomable levels of complexity, and at some point, we won't be able to tell real life from fiction. Although many of us are okay with that, artists have been very vocal against it. This form of entertainment feels like a threat, but for decades, artists have survived, and thrived, in spite of adversity. So, even with AI advancing, I don’t think we’ll ever not have artists.
